https://irjet.net/archives/V4/i7/IRJET-V4I756.pdf

1. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 327
Review Paper on Target Detection Using Ordered Statistical Method
Snehal V. Hiwase1, Prof. Veena A. Kulkarni2
1Master of Engg student Dept. of ECE, Pimpri Chinchwad College of engg Akurdi, Pune
2 Professor, Dept. of ECE, Pimpri Chinchwad College Of Engineering, Pune
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract -The main aim of constant falsealarmrate(CFAR)
in radar to detect the target in an unknown , time varying
clutter and noise environment while maintain the constant
value of false target detection. Conventional filter fail to set
the threshold value to detect the target due to the time-
varying nature of noise/clutter power. CFAR is a adaptive
algorithm that not only detect the radar target but also
maintain the increasing false target detection. Existing CFAR
algorithms are effective during certain environment only. So
new adaptive threshold technique is used to detect the target
called variability index constant false alarm rate(VI-CFAR)
which is combination of CA-CFAR, GO-CFAR and SO-CFAR. VI-
CFAR adapt itself to different algorithm depending upon
wether the environment is homogeneous, Non-homogeneous
and multiple target. Performance of VI-CFAR is better than
CA-CFAR, GO-CFAR and SO-CFAR. VI-CFAR give low loss
performance for homogeneous environment and robust
performance for non-homogeneous and multiple target
environment.
Key Words: Noise background environment, Constant
false alarm rate (CFAR), Variability index and mean
ratio.
1. INTRODUCTION
A stream of radar pulses is given to the duplexer .
Duplexer will on the transmitting antenna andpulseis
transmitted toward target .At that time Duplexer will
off the transmitting antenna and on the receiving
antenna . The return signal is allowed to display. This
received signal not only consist of noise but also
clutter. Clutter is unwanted noise signal whichisnotof
radar use . The example of clutter is the signal return
from mountain ,rain drop etc. Most of the technique
uses matched filter to fix the threshold but it fail due to
the presence of noise and clutter in received signal. In
radar target detection the probability of false alarm
rate is directly proportional to the amount of noise
present in received signal .So the small increasing in
total noise power will increase the probability of false
alarm to larger factor .
Fig -1: effect of noise power on PFA
Here it is seen from the graph that the desired
probability of false alarm is 100 but due to 8 dB
increase in noise power, probability of false alarm is
increased by a factor of 102 .Some time smaller target
produce signal which is weak in nature so may lost
completely in noise and clutter or in some cases the
amplitude of return echo signal is more than set
threshold value such that it is considered as a target.
Such cases is called as false target detection. So the
threshold must be stronger enough to detectthetarget
in presence of clutter and noise environment. This
increasing probability of false alarm will make data
processing equipment to saturate. So to maintain this
undesirable increasing in PFA ,and to maintain the
threshold as per input environment, an adaptive
threshold technique is needed. There are different
technique that not only maintain false target detection
but also reduces the effect of clutter and noise ,one of
them is Constant false alarm rate. The basic structure
of CFAR is shown in figure 2.
Fig -2: Method to detect target using CFAR algorithm

2. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 328
Here the input signal is any radar signal which may be
homogeneous, non-homogeneous and multiple target
which is given to square law detector. Which is then
given to shift register where the cell under test valueis
surrounded by guided cell and reference cell which is
given to CFAR algorithm . CFAR processor calculate
average value depending upon whether the
environment is homogeneous , non-homogeneousand
multiple target, which is multiplied with scaling factor
to generate threshold value . This generated threshold
value is compared with cell under test (CUT) value is
use to decide whether the target is present or not.
There are various method to detect the target which is
based on whether the environment is homogeneous ,
non-homogeneous andmultipletargetwhichisexplain
in next section.
2. METHODS
So to perform target detection variable threshold is need
which may adapt as per background. One such a adaptive
technique to detect the target is Cell averaging CFAR (CA-
CFAR).
Fig-3: Cell averaging CFAR (CA-CFAR) algorithm
2.1 Cell averaging CFAR (CA-CFAR)
In most of the case, the threshold calculation is based on the
amount of noise around cell under test . Target may be
present in cell under test or in reference cell. Cell averaging
is best to use in homogeneous environment . Homogeneous
environment is one which is free from clutter. In this case
averaging of cell is used to estimate noise backgroundwhich
is then multiplied with scaling factor is used to give
threshold which is compare with cell under test value use to
decided whether the target is present in cell under test or
not. Adaptive threshold for target cell averaging CFAR is
given by
Q = U+V
Whenever there is a transition from clear to clutter
environment is not smooth then performance of Cell
averaging CFAR is decreases. This method is not applicable
to detect the target when another interfering target is their.
To overcome this disadvantage next method is proposed.
2.2 Greatest of CFAR (GO-CFAR)
Whenever there is a transition from clear to clutter is not
relatively smooth then two cases arises In first case cell
under test is in clear region and number of reference cell in
clutter region then adaptive thresholdvalueisincreasedand
probability of false alarm decreasesandinsecondcasewhen
cell under test is in clutter region and number of reference
cell is in clear environment then adaptive threshold value is
decrease and probability of false alarm increases[8]. For
such cases CA-CFAR fail to give good performance . To
overcome this disadvantage GO-CFAR is proposed . Herethe
adaptive threshold value is nothing but maximum value
between two reference window which is given as
Q = Max(U,V)
But performance of GO-CFAR decrease in presence of
another interfering target in same reference window or its
performance decreases in case of Multiple target. To
overcome this disadvantage next method is proposed.
2.3 Smallest of CFAR (SO-CFAR )
SO-CFAR uses the same circuit as that of CA-CFAR to detect
the target only the difference is instead of average or
maximum, SO-CFAR uses smallest value. Insteadofchoosing
the window with higher noise level it chooses the window
with lower noise. The adaptive threshold value for SO-CFAR
is given as
Q =Min (U,V)
From above discussed method it is clear that all this method
is applicable to particular environment only. As the
environment changes performance of different algorithm
decreases. So to overcome this disadvantage new algorithm
is presented called Variability Index constant false alarm
rate (VI-CFAR)
2.4 Variability index constant false alarm rate (VI-CFAR)
Variability index constant false alarm rate (VI-CFAR) is an
adaptive algorithm composedofCA-CFAR,GO-CFARandSO-
CFAR. Depending upon whether the environment is
homogeneous , non-homogeneous and multiple target , it
will adapt itself different algorithm. VI-CFAR select the
threshold value based on background estimation algorithm.
Background estimation algorithm work based onVariability
index (VI) and Mean ratio (MR) values.
[I]. VARIABILITY INDEX STATISTICS

3. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 329
Variability index (VI) is estimation of probability density
function (pdf). Pdf of Variability index is independ on noise
power in homogeneous environment but affected during
interfering target and reverberation within reference
window. For each leading and lagging window , VI is
calculated and is given by
VI= 1+
Which is compare with threshold KVI to decide whether the
environment is homogeneous or non-homogeneous.
[II] C. Mean Ratio Statistics
Ratio of mean value of leading reference windowAtomean
value of lagging reference window B is nothing but Mean
Ratio and is given as follow
MR =
Comparison of calculated MR and threshold KMR (1.806) [2]
is decided whether the population mean in the leading
reference window and lagging reference window are same
or different.
[III]. VI-CFAR Threshold Calculation
Adaptive threshold of VI-CFAR is based on the result of the
VI hypothesis test and the MR hypothesis test. These result
decide which window is used to estimatebackgroundpower
and a constant multiplier which is usedtocalculate adaptive
threshold. The background multiplier CN and CN/2 dependon
which reference window is selected to estimate background
power. Following table is used to give which method is used
to calculate the value of CN and CN/2.
Table-1: DIFFERENT VI-CFAR MODES OF OPERATION [2]
Leading
window
A
variable
Lagging
window
B
variable
Different
mean
VI-CFAR
Adaptive
threshold
CFAR
Methods
No No No CN.∑AB CA-CFAR
No No Yes CN/2. max(∑A
,∑B )
GO-CFAR
Yes No - CN/2 . ∑B CA-CFAR
No Yes - CN/2 . ∑A CA-CFAR
Yes Yes - CN/2. min(∑A
,∑B )
SO-CFAR
Where CN = and CN/2 =
3. RESULT
While considering homogeneous environment i.e black
sample image shown in figure 4 when pass to the
algorithm we obtain the probability of false alarm value 0
which is shown in figure 4.
Fig 4-: Black background image 1 sample and obtain PFA
value.
Black Background sample image with white target
highlighted is shown in figure 5 when pass through
algorithm we obtain probability of false alarm value12.This
means the algorithm is able to detect the presenceofTarget.
Fig-5: Black background image2 sample with white target
highlighted and obtain PFA value.
Figure 6 shows result of CA-CFAR, OS-CFAR, GO-CFAR, SO-
CFAR when Black Background sample image with white
target highlighted is pass through algorithm.

4. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 07 | July -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 330
Fig-6: Black background image 2 sample and Result of Cell
average-CFAR, Ordered statistic-CFAR , Greatest of-CFAR,
Smallest of-CFAR .
3. CONCLUSIONS
Every method is limited to particular environmentonlysuch
as CA-CFAR performance is good for homogeneous
environment Itsperformancedecrease whenmovingtoward
clutterenvironmenttoovercomethisdisadvantageGO-CFAR
is presented but it is limited to non-homogeneous
environment so S0-CFAR is presented to detect multiple
target. But none of method detect target in all environment
so VI-CFAR is presented which give good performance in all
three environment.
ACKNOWLEDGEMENT
It would not have been possible to complete this project
work without the kind support and help of many individuals
and organizations. I would like to extend my sincere thanks
to them. I would like to thank my guide , Prof. V. A. Kulkarni,
Electronics & Telecommunication Department, for his
encouragement and support.
REFERENCES
[1]. Vladimir G. Galushko, O. Ya. Usikov, “Analysis of the CA
CFAR Algorithm as Applied to Detection of Stationary
Gaussian Signals Against a Normal Noise Background”,IEEE
Transactions on Institute of Radiophysics and
Electronics of National Academy of Sciences of Ukraine
Kharkiv, 24 June 2016
[2] Jong-Woo Shin, Young-Kwang Seo,[et.al], “Modified
Vriability-Index CFAR Detection Robust to Heterogeneous
Environment”, International Conference on Systems and
Electronic Engineering , Phuket (Thailand), (ICSEE'2012)
December 18-19, 2012 .
[3]. Amit Kumar VermaI, “Variability Index Constant False
Alarm Rate (VI-CFAR) for Sonar Target Detection”, IEEE-
International Conference on Signal processing,
Communications and Networking Madras Institute of
Technology, Anna University Chennai India, pp'38-141, Jan
4-6, 2008.
[4].S. E. Thompson [et al] “Automatic Censoring CFAR
Detector Based on Ordered Data Variability for non
homogenous Environments.”IEEE Transactions On Radar
And Electronic ,February 2010.
[5].Hermann Rohling[et.al] ”Radar CFAR Thresholding in
clutter and Multiple Target Situations “IEEE Transactions
Aerospace And Electronic Systems Vol. Ase- 19, NO. 4 July
2002.
[6].Nadav Levanon [et.al] “Detection Loss Due to Interfering
Target in Ordered Statistic CFAR”, IEEE Transactions On
Aerospace And ElectronicSystemsVol.24.NO.6NOVEMBER
2005.
[7] Phillip E Pace, “False Alarm Analysis of the Envelope
Detection GO-CFAR Processor”, IEEE Transactions On
Aerospace And Electronic Systems Vol. 30, No. 3 July 1994
[8] Mourud krket and Pramod K. Vershney “On Adaptive
Cell-Averaging Cfar Radar Signal Detection”, Rome Air
Development Center Air Force Systems
Command Griffiss Air Force Base, 7 June 1988.